Nvidia's revenue and profit in the second quarter doubled year-on-year, and its CFO released a new guidance of 70% growth.

After US market hours on Wednesday, NVIDIA released its financial results for the second quarter of fiscal year 2027. The company's two core profitability indicators, quarterly revenue and net profit, both significantly outperformed market expectations. Meanwhile, its Chief Financial Officer projected a 70% revenue growth for fiscal year 2028, compared to the previous market consensus estimate of 45%.

The stock price fluctuated after the earnings release: it briefly dipped first, then rallied rapidly boosted by the management's 70% revenue growth guidance for fiscal year 2028, with a maximum post-market increase of 4.7% at one point.

The pessimistic sentiment brought by the previous seven consecutive trading days of decline, during which the stock price fell by 7.5% cumulatively, was significantly eased. This seven-session losing streak marked the longest consecutive decline in four years since September 2022, and the price was down more than 10% from the all-time closing high of $235.74 set on May 14 this year.

Boosted by NVIDIA's strong performance outlook, sectors including memory, optical communication, and AI cloud services collectively surged after hours: SK Hynix's ADR rose by more than 4%, Applied Optoelectronics climbed over 4%, CoreWeave increased by over 5%, and Nebius jumped more than 6%.

Earnings data shows that NVIDIA recorded total revenue of $96.22 billion this quarter, a sharp 106% increase from the same period last year and an 18% rise compared to the first quarter. Prior to this, the market consensus revenue expectation aggregated by LSEG was about $92.17 billion. The actual revenue scale exceeded the expectation by more than $4 billion, and there was no sign of the slowdown in growth momentum that the market had previously worried about.

On the profitability side, the company's net profit in the second quarter reached $53.95 billion, doubling year on year, and the profit expansion pace kept pace with the high revenue growth trend.

In terms of profitability, both GAAP and non-GAAP gross margins of NVIDIA in the second quarter remained stable at 75.0%, which is in the high range of the industry. The adjusted earnings per share was $2.22, up 120% year on year, higher than the market expectation of $2.10, and the earnings quality exceeded the capital market's forecast.

Breaking down business segments, the data center business remains the absolute core pillar driving NVIDIA's performance growth. Revenue of the data center segment in the second quarter reached $89 billion, up 117% year on year and 18% quarter on quarter, accounting for about 92.5% of the company's total revenue in the current period.

In the data center customer structure, hyper-scale cloud service providers contributed $48.71 billion in revenue, up 102% year on year; AI cloud vendors, industrial and physical enterprise customers brought in $40.3 billion in revenue, with a year-on-year growth rate as high as 138%. The growth rate of AI computing power procurement demand on the enterprise side has exceeded that of large cloud vendors, and the downstream computing power demand landscape is further expanding.

The revenue of other segments including consumer gaming graphics card business, professional visualization, and automotive business accounts for a relatively small proportion, and the overall operation remains stable without obvious fluctuations. From the signal on the demand side, computing power procurement orders from enterprises and AI startup laboratories continue to be released, becoming the second growth curve following large cloud vendors.

On the earnings call, NVIDIA's Chief Financial Officer Colette Kress said NVIDIA's performance growth rate will accelerate next year, and revenue is expected to grow by 70% in fiscal year 2028. She also stated that "even at our current scale, demand is still accelerating. Customer forecasts show that our growth next year will double."

Kress also added that the real demand growth rate of computing power procurement measured by downstream customers at present is even expected to approach 100%, but limited by the production capacity constraints in links such as the chip supply chain and memory components, the scale of products that NVIDIA can stably deliver at this stage can only support revenue growth of about 70%.

This statement sends a signal that the short-term ceiling of the company's performance does not come from insufficient market demand, but is restricted by the supply bottleneck of the upstream supply chain.

NVIDIA CEO Jensen Huang said that if there were no supply restrictions, the company's performance outlook for fiscal year 2028 could be "much higher".

For the operating situation in the third quarter of fiscal year 2027, NVIDIA also gave quarterly revenue guidance, expecting the quarterly revenue scale to be about $108 billion with a floating range of 2%. The GAAP gross margin expectation for the third quarter is around 74%, with a fluctuation of 50 basis points, slightly lower than the 75% level in the second quarter, which confirms the prediction of the impact brought by the rise in memory costs. This figure is lower than the 75% expected by analysts, which caused NVIDIA's stock price to dip by more than 4% at one point after hours.

NVIDIA CEO Jensen Huang once again gave an optimistic judgment on the long-term prospect of the AI industry on the earnings call. He said: "Artificial intelligence has reached an inflection point. AI is generating practical utility. Tokens are productive and can generate revenue, and computing power itself is income. At present, the demand for computing power is accelerating, and we are in a golden period for the development of new AI laboratories and startups. Many cutting-edge laboratories are simultaneously expanding their computing power scale."

This earnings report of NVIDIA that exceeded expectations has also injected a shot in the arm into the entire computing power industry chain. Since the AI wave started, NVIDIA has maintained a doubling-level revenue growth rate for consecutive quarters. However, behind the rapid growth, there are still risk factors worthy of the market's vigilance.

On the one hand, competitors such as AMD and Google are continuously launching new generation of computing power products, bringing increasing competitive pressure. At the same time, the global shortage of memory chips has not been alleviated, which has also caused NVIDIA's memory procurement costs to rise sharply, both of which may become disturbing factors for subsequent performance growth.

Data from Morgan Stanley shows that the selling price of NVIDIA's new flagship AI rack product Vera Rubin is about $7.8 million, and the memory cost has soared from about $370,000 to $2 million, a surge of 435%, accounting for 26% of the total machine cost, up from less than 10% of the previous generation.

In addition, NVIDIA has long made equity investments in core customers such as OpenAI and Anthropic, blurring the boundary between suppliers and customers, and some institutions have begun to examine the potential risk points behind NVIDIA's revenue data.

This article is from the WeChat official account "Jiemian News", Author: Shen Xiaoge, 36Kr publishes this content with authorization.